Paper Title
Optimized Deep Learning Model for Lung Cancer Diagnosis Leveraging Efficientnet-B7 and Transfer Learning

Abstract
Lung cancer remains a leading cause of cancer mortality, in large part due to diagnostic delays and the challenge of persistent manual interpretation of CTscans. In this work, a deep learning-based framework designed for automatic classification of lung CT images into benign, malignant, and normal classes. EfficientNet-B7 is used as the feature extractor and is fine-tuned using transfer learning to minimize the requirement for large training data. The model is trained using the IQ-OTHNCCD dataset and evaluated using standard metrics and visualization tools, such as confusion matrices and classification reports. The proposed method shows high reliability in discriminating among the three classes, suggesting its ability to assist radiologists in performing fast and reliable clinical assessment. Generally, the results indicate the ability of transfer-learning-based deep models to assist in early lung cancer screening and decision making in medical settings. Keywords - EfficientNet-B7, CT scans, Transfer Learning, IQ-OTHNCCD Dataset, Benign, Malignant, Normal, Medical Imaging, Automated Diagnosis.